|
All lenders’ business models must either be able to hold loan product indefinitely or insure that its exit strategy is viable under all circumstances short of a state of armed invasion – preferably both. Prudence on both sides of this paradigm is akin to having more than one source of repayment on the loan (cash flow, collateral value, borrower recourse and guarantors).
Determining how much of the collateral value to advance is fundamental to lending. It is, in fact, the basis of all collateral lending and underwriting activities. Haircutting more specifically refers to the way the CP (“Commercial Paper”) conduits and other “warehouse” lenders, the Federal Reserve Banks and the FHLBs (“Federal Home Loan Banks”) decide how much to advance against pledged assets. They are all essentially making a judgment about how much the value of an asset can fall (price volatility) during the expected funding period. All strong lenders necessarily perform the same analysis in performing their credit function The resulting “haircut” each lender imposes on how much of the value will be lent against has many names, such as “subordination”, “loan to value”, “skin in the game” and “advance rate”.
Nature of the Collateral Determines Value Volatility
Collateral value volatility is determined first and foremost by what the assets are and the market for them. At the most liquid end of the spectrum, US Government bonds may require very small margins. While they can suffer value declines due to changes in market yields, they will always pay in full at maturity. So the funding horizon matters in this context, but not the credit. If the collateral is stocks, a sudden break in the market may, in a few hours - and before there is an opportunity to sell - turn a comfortable margin into an actual loss. However the equity markets are generally extremely liquid and price discovery operates continuously. Where the collateral is loans of a variety of types and of varying quality, and price discovery is poor, the only way to determine appropriate margins is to examine potential value volatility. Scenario simulations are how you discover the appropriate haircut / margin for price volatility.
Analogs – Marginable Securities
Tests for Marginable Securities: It is common in margining equities for institutional clients to require that the value of the collateral stock always be 20% more than the amount advanced.
A second rule that discriminates against low-priced stocks not paying a dividend is also applied by many institutions – the ten point (some institutions use 15 points) rule. That is, the amount loaned must be at least 10 points less than the value of the collateral. For example, in a mixed portfolio with 2000 stocks and a margin loan of $100,000 the margin must be at least $20,000 and the collateral value must be at least $120,000. Divide the margin by the shares. The result must be at least $10.
Currently, Charles Schwab requires a 30% minimum margin, with a 50% initial margin at purchase for retail buyers.
Back in the Old Days of Mortgages: There was a time when no bank would make a loan to an individual who didn’t have the ability to put down 20% of the purchase price (from Joseph Edward Morton, Urban Mortgage Lending: Comparative Markets and Experience. Princeton, NJ: Princeton University Press, 1956) and loans often amortized fully in 15 years or ballooned in 10. Government market interventions since 1934 have steadily driven down those requirements to the great benefit of many, but at a cost: The lending system has no equity cushion when collateral values fall as they have recently. The system also has increased risk because borrowers have less “skin in the game”. Fresh equity makes a difference, particularly in commercial lending.
Because Loans are Less Liquid, Value Must be Ascertained by Pricing Different Scenarios: MIAC sees the value of a pool of loans as a function of the offering (size, quality, etc.) in the context of the market at any given time (capacity of buyers and types of buyers – market “saturation” factors – as well as the market perception about seller need and asset quality). A seller’s ability to hold loans and time sales has a dramatic impact on sale value realized over time. No one pricing scenario is always the right one.
Components of a Haircut
The components of a haircut should include all the costs that a buyer would impute, plus the costs to execute a sale. MIAC sees the primary components as:
- Credit Loss (OTTI)
- Yield (Market Discount)
- Volatility Adjustment (potential for deterioration – often called “Liquidity”)
- Costs of Liquidation
The costs are a small part of the whole. They tend to grow larger in percentage terms for smaller or more complex portfolios because the base costs involved reach a minimum size. Volatility, defined as how much value can deteriorate, is the big unknown.
Possible Valuation Scenarios
There are a variety of possible scenarios for valuation purposes. In the market today, the possible scenarios really can be boiled down to a handful of variants.
FAS 157 Fair Market Value: This methodology provides a value based on the premise that there is a willing buyer and a willing seller. Implicit in that assumption is the corollary that willing sellers don’t need to sell – particularly in distressed markets. This scenario carries an accurate assumption of future losses and a realistic assumption about market yields for the product.
Held to Maturity Value: This scenario uses the FAS 157 credit loss assumptions, but assumes the actual yields on the assets are satisfactory to the originator.
Commercial Paper Conduit “Haircut” Methodology: This is the general methodology employed by the commercial paper conduits that are in the business of “warehousing” loans prior to sale or securitization – up to 364 days. Their “haircut” has to support a commercial paper rating for them to maintain their funding. The models stress key variables to arrive at higher PODs and/or LGDs per an approved rating agency criteria. Both product-specific rating agency teams and CP teams must pass on the methodology employed to support ratings.
Distressed Market Value: This is the price a seller who MUST sell (and frequently has poor information on the loans) must sell at in either good or bad markets. Regardless of market tone, distressed pricing is lower than fair market value, but how low depends on the state of the market at the time of sale. Both market yield and credit loss assumptions are elevated due to changing expectations of potential buyers and because collateral information is often poor in such situations.
No One Valuation Scenario is Perfect
MIAC believes that no one single valuation scenario is “correct” for a variety of reasons. These reasons include:
> Seller Deterioration: Changes in the condition of the seller will change the expectations of the market and thus loan value. A drastically weaker potential seller rapidly becomes viewed as distressed, bringing with that a shift in the assumptions a potential buyer will employ. At bottom, buyers assume the loans are “scratch and dent” because the seller is failing.
> Changing Buyer Universe: The universe of potential buyers constantly changes. Even where distress is not a factor, market participants who have the necessary amount of cash to invest will shift as product pays off, and as new product is added to the market and then absorbed. It also must be remembered that the “vultures” aren’t the only ones who seek bargains. The main street consumer and the main street banker also seek bargains.
> Limited Conditions for FAS 157: FAS 157 value is ONLY true when buyers and sellers agree. While this does happen, in depressed market conditions it doesn’t happen as often – certainly not on demand. Also, the value is based on assumptions about reasonableness. Sometimes sale conditions are not reasonable.
> Conduit Method Can be Over or Under: The Conduit valuation methodology can be too low (hot markets) and it can be too high (depressed markets). It has, however, the advantage of consistency and a measure of conservatism under “normal” market conditions. It also can be low or high compared to the market at any time, depending on the rating agency grade level to which the collateral has been priced. The methodology can be set to generate a haircut designed to support most any grade level from “AAA” all the way down to junk bond status. The methodology is intended to cover depressed market conditions, but the derived assumptions are essentially specific levels based on statistical averages covering past market downturns. Because of continuous changes in the universe of offered product and buyers available, the level obtained can vary considerably from a “market” level at any one time.
> Distressed Market Pricing: This methodology will produce the lowest price whenever markets are depressed. In a hot market, however, it could actually produce a price above the Conduit model AA or AAA value. The results are market-specific and asset specific. They have no clear consistency over time.
What Are the Best Options?
Guess Conservative: One can be aggressive, conservative or very conservative. The table below is a variant of the one shown earlier in this article, only this one shows final price less a volatility adjustment based on the Distressed rather than FMV (“Fair Market Value”) price. This approach factors in a cushion against a worst-case deterioration of the borrower.
Preserve Value and Monitor Data: Failing banks have missed out because when their markets changed they were caught by surprise. In our analog, no stock broker would margin stock where daily market values are not available. Most small banks do not:
> Require updated financials and rent rolls on businesses (i.e., Commercial and Industrial (C&I loans)) or commercial real estate (CRE loans) that must generate cash for payback.
> Obtain updated credit bureau data on consumer-borrowers
> Obtain updated appraisals or AVMs (Automated Valuation Models) of any kind on large credit exposures, such as ADC (Acquistions, Development and Construction) loans and CRE loans. The maxim “You can’t get hurt if you lend on dirt” has been proven wrong. Cash flow pays you back – such as unit sales on ADC loan projects. Reputable loan brokers want this data developed to sell loan assets.
Re-Run Multiple Value Scenarios Regularly: Have loans valued on more than a FMV basis regularly. Know how much value will decline if the unthinkable happens. Running valuations under multiple scenarios that encompass the possible things that can go wrong is the ONLY way to correctly build a haircut.
Conclusion
The scenario analysis needed for good ALM, bank stress testing and liquidity planning are the keys to correctly ascertaining value volatility. These tools are a financial institution’s best way of insuring that credit standards for “haircuts” (LTVs, etc) meet or exceed prudent standards for all product types. MIAC does this type of analysis regularly and can help you see the alternatives, and what you need to do in order to preserve and enhance franchise value.
|